Hacker News Reader

What is holding Intel back, there?

Engineering chops?

AMD and nVidia already patented too much of the good stuff?

Too much existing code optimized for AMD and nVidia quirks?

Extreme outsider perspective but they seemed like dilettantes. They'd dip their toe into doing GPUs and then cancel the project every couple years.

A few weeks ago Gelsinger even commented he saw "less need for discrete graphics in the market going forward" - just seemed like a very Intel thing to say

As a fellow outsider, that makes a lot of sense to me. Even if Intel started outperforming nVidia today... who would want to put serious effort into making their stuff work with Intel hardware in a market they're likely to pull out of at any moment?

The early stuff (Larabee, KNL, KNC etc.) was hamstrung by x86. Wrong architecture for the types of things GPUs are good at. Their igpus have generally been good but that's not competing in the compute segment. Then they acquired a few startups (Nervana, Habana) that didn't really work out either. And now they finally have a discrete GPU lineup that is making slow progress. We'll see.

Reply on news.ycombinator.com